Viewing profile — nknutalapati

Hi HN — I've been experimenting with ways to make AI agent actions auditable and enforceable at runtime. This project has two parts. 1. LICITRA-MMR — An append-only audit log using…

Localhost-only dashboard with SQLite audit is a good default for self-hosted. Same question I keep asking these proxy projects: if the SQLite log is the evidence layer, what happen…

Runtime guarding at the tool execution layer is the right enforcement point. One thing I'd push further: the audit trail — is it append-only with integrity guarantees, or a standar…

Solid proxy architecture. The deny-by-default YAML policy engine is the right call. One question on the audit side: decisions are logged to SQLite — is that log tamper-evident? If …